Github user drewrobb commented on a diff in the pull request:
https://github.com/apache/spark/pull/23062#discussion_r234861629
--- Diff:
sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/ScalaReflectionSuite.scala
---
@@ -109,6 +109,35 @@ object TestingUDT {
}
}
+/** An example derived from Twitter/Scrooge codegen for thrift */
+object ScroogeLikeExample {
+ def apply(x: Int): ScroogeLikeExample = new Immutable(x)
+
+ def unapply(_item: ScroogeLikeExample): Option[Int] = Some(_item.x)
+
+ class Immutable(val x: Int) extends ScroogeLikeExample
+}
+
+trait ScroogeLikeExample extends Product1[Int] with Serializable {
+ import ScroogeLikeExample._
+
+ def x: Int
+
+ override def _1: Int = x
+
+ def copy(x: Int = this.x): ScroogeLikeExample = new Immutable(x)
+
+ override def canEqual(other: Any): Boolean =
other.isInstanceOf[ScroogeLikeExample]
+
+ private def _equals(x: ScroogeLikeExample, y: ScroogeLikeExample):
Boolean =
+ x.productArity == y.productArity &&
--- End diff --
I'm worried about changing the tests to use a concrete subtype, because the
reflection calls might behave differently in that case either now or later on.
I simplified a little more. `canEqual` is necessary to implement product.
`equals` is necessary or tests will not pass (it will check object pointer
equality), and `hashCode` is needed for scalastyle to pass since `equals` is
necessary.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
